Aspirin delays the onset of hypertensive disorders of pregnancy among nulliparous pregnant women: A secondary analysis of the ASPIRIN trial

To assess the impact of low‐dose aspirin (LDA) starting in early pregnancy on delaying preterm hypertensive disorders of pregnancy.


| I N TRODUC T ION
Hypertensive disorders of pregnancy (HDP) remain a primary contributor to maternal and perinatal death and result in longterm health risks for both the mother and child.The risk of such complications is greater when the disease is severe and of early onset resulting in preterm birth.Estimates suggest that HDP is responsible for approximately one-sixth of all preterm births, and is frequently associated with fetal growth restriction. 1,2Preterm birth and fetal growth restriction are known to have lifelong consequences for the child, including a higher risk of neurodevelopmental delay, respiratory disorders, renal dysfunction, insulin resistance, obesity, hypertension, cardiovascular disease and impaired work capacity. 3 Additionally, HDP results in adverse maternal outcomes; for example, mothers affected by pre-eclampsia are two to five times more likely to develop cardiovascular diseases including systemic hypertension and cerebrovascular disease in the future when compared with women who were not hypertensive during pregnancy. 4eterogeneous studies have evaluated the possible benefit of low-dose aspirin (50-150 mg/day) in pregnancy to minimise the risk of HDP, with large variations in the included population, aspirin dosage, gestational age at initiation and disease definition. 5][8][9][10] Aspirin use in pregnancy may be an ideal intervention because of its wide availability, relative inexpense, and few (if any) safety concerns for the mother and fetus. 11onetheless, the effects of LDA on overall HDP are modest, which may reflect the heterogeneous aetiologies of the disease.Increasingly, the literature has suggested that early HDP (<34 weeks) is secondary to a failure to achieve deep placentation while term pre-eclampsia reflects placental senescence.Though this distinction in the literature has emerged, few primary trials and most of the meta-analyses have not examined HDP in the framework that the role of LDA is not to prevent all HDP but rather to delay it.Recently, a non-prespecified analysis of the ASPRE trial 12 demonstrated that women who were at high risk for HDP by clinical, ultrasound and biochemical

Setting:
The study was conducted among women in the Global Network for Women's and Children's Health's Maternal and Newborn Health Registry (MNHR) clusters, a prospective, population-based study in Kenya, Zambia, the Democratic Republic of the Congo (DRC), Pakistan, India (two sites-Belagavi and Nagpur) and Guatemala.Population: Nulliparous singleton pregnancies between 6 +0 weeks and 13 +6 weeks in six low-middle income countries (Democratic Republic of Congo, Guatemala, India, Kenya, Pakistan, Zambia) enrolled in the ASPIRIN Trial.Methods: We compared the incidence of HDP at delivery at three gestational age periods (<28, <34 and <37 weeks) between women who were randomised to aspirin or placebo.Women were included if they were randomised and had an outcome at or beyond 20 weeks (Modified Intent to Treat).Main Outcome Measures: Our primary outcome was pregnancies with HDP associated with preterm delivery (HDP@delivery) before <28, <34 and <37 weeks.Secondary outcomes included small for gestational age (SGA) <10th percentile, <5th percentile, and perinatal mortality.Results: Among the 11 976 pregnancies, LDA did not significantly lower HDP@delivery <28 weeks (relative risk [RR] 0.18, 95% confidence interval [CI] 0.02-1.52);however, it did lower HDP@delivery <34 weeks (RR 0.37, 95% CI 0.17-0.81)and HDP@delivery <37 weeks (RR 0.66, 95% CI 0.49-0.90).The overall rate of HDP did not differ between the two groups (RR 1.08, 95% CI 0.94-1.25).Among those pregnancies who had HDP, SGA <10th percentile was reduced (RR 0.81, 95% CI 0.67-0.99),though SGA <5th percentile was not (RR 0.84, 95% CI 0.64-1.09).Similarly, perinatal mortality among pregnancies with HDP occurred less frequently (RR 0.55, 95% CI 0.33-0.92) in those receiving LDA.Pregnancies randomised to LDA delivered later with HDP compared with those receiving placebo (median gestational age 38.5 weeks vs. 37.9 weeks; p = 0.022).Conclusions: In this secondary analysis of a study of low-risk nulliparous singleton pregnancies, early administration of LDA resulted in lower rates of preterm HDP and delivery before 34 and 37 weeks but not in the overall rate of HDP.These results suggest that LDA works in part by delaying HDP.

Funding information
Eunice Kennedy Shriver National Institute of Child Health and Human Development, USA markers randomised to aspirin at 150 mg had a 4.4-week delay in delivery with HDP. 13 Given that this finding has not been evaluated in other populations, we undertook this secondary analysis to assess the impact of aspirin administration as compared with placebo on the gestational age (GA) at delivery in women with HDP in the recently completed 'Aspirin Supplementation for Pregnancy Indicated Risk Reduction In Nulliparous Women (ASPIRIN)' trial, which randomised 11 976 nulliparous women in six low-middle income countries to either LDA (81 mg daily) or identical placebo between 6 +0 weeks and 13 +6 weeks through 36 weeks' GA of pregnancy.

| M ET HODS
We performed a non-prespecified secondary analysis of the ASPIRIN trial examining the impact of LDA on the gestational age at delivery of nulliparous pregnant women with HDP.The ASPIRIN trial enrolled nulliparous women from seven community sites in six low-and middle-income countries (LMIC) (two sites in India, and one site each in Pakistan, Zambia, the Democratic Republic of the Congo (DRC), Guatemala and Kenya) between March, 2016 and April, 2019.The trial was led by the Global Network (GN) for Women's and Children's Health Research (Eunice Kennedy Shriver National Institute of Child Health and Human Development [NICHD], Bethesda, MD, USA).Each site enrolled women from primary healthcare centres and hospital-based clinics.The trial protocol has been published previously. 14The study protocol was approved by the pertinent ethics committees and regulatory agencies of each participating country and the ethics committees of the US collaborating institutes and that of the RTI International.Written informed consent for participation in the trial was obtained from the women or the guardians/parents of minors, by study staff at each site.
Nulliparous pregnant women aged between 18 and 40 years (minors aged ≥14 years were enrolled when permitted by individual ethics boards in the DRC, Kenya and Zambia) with a singleton gestation between 6 +0/7 weeks and 13 +6/7 weeks without evidence of a fetal anomaly identified by ultrasound were deemed eligible.Women with a history of allergy or contraindication to aspirin; previous consumption of aspirin for more than 7 days during the pregnancy; multiple gestations; more than two first-trimester losses; any other medical condition that might be considered a contraindication to inclusion in the study (e.g.diabetes and hypertension) were excluded at the discretion of the investigators. 15Women were also excluded if their blood pressure was >140/90 mm Hg or haemoglobin was <7.0 g/dL at the time of enrolment.The crown-rump length and date of the last menstrual period (LMP) was used to assess the gestational age with a mobile application in accordance with ACOG guidance. 16ligible women who gave their consent were randomly assigned (1:1) to receive 81 mg of aspirin or placebo until 36 weeks' gestation as detailed in the primary study.Medication adherence was noted to be high with 84.9% of women taking 90% of doses as assessed by pill counts. 17Blood pressure was assessed by field staff who had been trained in the appropriate technique of taking blood pressures using a standardised blood pressure cuff at regular intervals: 16-20 weeks, 28-30 weeks, 34 weeks, and then every 2 weeks from 34 weeks until delivery.
Our primary outcome was the gestational epoch (<28, <34 and <37 weeks) at which a pregnancy was delivered with HDP.Women were classified as having HDP if they developed BP >140/90 mmHg after 20 weeks' GA on at least two All cases were reviewed by a committee blinded to treatment allocation.As the availability and routine assessment of proteinuria differed between sites, this was not included in the definition of HDP. 18Care and management of women with HDP was at the discretion of the local care delivery system.Following delivery, we obtained maternal and neonatal outcomes up to 42 days using the Global Network Maternal Newborn Health Registry. 19,20

| Statistical analyses
To assess the onset of HDP and associated delivery we chose to examine three gestational age epochs (<28, <34 and <37 weeks).Data analyses were performed using SAS software (version 9.4) and graphics were produced using 'R' (version 4.0.5).Consistent with the primary study, women who delivered at ≥20 weeks' gestation with known status of HDP were included in this analysis (modified intent to treat, mITT).Descriptive statistics of socio-demographic, obstetric and clinical characteristics of women with HDP were reported by treatment group.Gestational age at delivery with HDP was compared visually between treatment and placebo groups using dot plots.
For each of the primary and secondary maternal and fetal outcomes, relative risks (RR) and 95% confidence intervals (CI) comparing aspirin with placebo were calculated using a generalised linear model with a binomial distribution and log link adjusted for site.If a model did not converge, RR and 95% CI were calculated using Mantel-Haenszel estimates of common relative risk stratified by site.In the case of zero cell counts, a logit estimator of the common relative risk which adds 0.5 to each cell frequency in the stratum is reported.The estimated mean difference and 95% CI for gestational age at delivery was obtained from a generalised linear model adjusting for site and treatment group.Gestational age at delivery was log-transformed and the estimates then back-transformed.
Quantile regression was performed on the subgroup of women with HDP using a log transformation of GA at delivery and including terms for treatment and site.The model estimates were then back-transformed and plotted by quantile and treatment group.

| R E SU LTS
Among study participants randomised, 11 558 (96.5%) were included in the mITT analysis; 11 544 (99.9%) with known HDP status have been included in this analysis (Figure 1).Maternal characteristics among women with HDP are described in Table 1.The baseline characteristics of the study population across the groups were similar with respect to age, education, gestational age at randomisation, baseline haemoglobin and anthropometric measures.More than 95% of the women had 3+ antenatal visits among both the groups.Overall, the Asian (India and Pakistan) and Central American (Guatemala) sites differed from the African (DRC, Kenya and Zambia) sites in the proportion of HDP cases.
In the primary analysis of the ASPIRIN Trial, there was no difference in the overall rate of HDP between women receiving LDA and placebo (LDA 6.1% vs. placebo 5.6%; RR 1.08, 95% CI 0.94-1.25). 17Women who had been randomised LDA and developed HDP were less likely to deliver before 34 weeks (RR 0.37, 95% 0.17-0.81)or before 37 weeks (RR 0.66, 95% CI 0.49-0.90)than were those treated with placebo.Women randomised to LDA who developed HDP were less likely to deliver before 28 weeks (RR 0.18, 95% CI 0.02-1.52)but this difference did not achieve statistical significance (Table 2).No differences in other maternal outcomes among women with HDP receiving LDA were evident with regard to vaginal bleeding, antepartum haemorrhage, postpartum haemorrhage and maternal mortality through 42 days post-delivery.
Assessment of fetal and neonatal outcomes among women with HDP revealed that the incidence of low birthweight (LBW; <2500 g) was less among the aspirin group than the placebo group (RR 0.71, 95% CI 0.57-0.87).Aspirin use was significantly associated with a reduced incidence of extremely LBW (<1500 g; RR 0.13, 95% CI 0.03-0.55).Babies born to women with HDP who received LDA had lower rates of small for gestational age (SGA) <10th percentile (RR 0.81, 95% CI 0.67-0.99)but not <5th percentile (RR 0.84, 95% CI 0.64-1.09).Fetal losses and stillbirth were significantly lower among women with HDP who received LDA as compared with placebo.
The distribution of gestational age at delivery for women with HDP in the aspirin and placebo groups is shown in the dot plot (Figure 2).This graph demonstrates that women randomised to LDA who later developed HDP appear to deliver at a later gestational age than do women who received the placebo.This suggests that LDA may delay the onset of HDP at delivery.The mean GA at delivery was higher for women randomised to LDA than to placebo (LDA 38.5 weeks vs. placebo 37.9 weeks; RR 1.02, 95% CI 1.01-1.03).In addition, the overall median GA at delivery was higher in the LDA group than in the placebo group (LDA 38.7 weeks vs. placebo 38.4 weeks; p = 0.022).In a post-hoc analysis of women with HDP, quantiles of gestational age at delivery were observed to be higher in women who received LDA than in women who did not receive LDA up to the 0.35 quantile (see Figure 3 for the estimated quantiles in both groups, adjusted for site).There was no evidence of a difference between these two groups in quantiles >0.35, suggesting that LDA may benefit women with HDP who deliver in earlier gestational windows.Figure 4 depicts the survival analyses for time until delivery with HDP for both aspirin and placebo groups.Participants were censored at time of delivery if there is no HDP.The inset shows the same data on an enlarged yaxis.Hazard ratios and 95% CI for HDP comparing aspirin with placebo within three GA time periods (early preterm, preterm and term) are calculated using a Cox proportional hazard model with a time-varying treatment arm.The probability of women with HDP delivering at an early GA had a smaller hazard ratio with later GA at delivery.The hazard ratios (95% CI) calculated for different GA ranges are as follows: for 20-33 weeks, 0.69 (0.49-0.97); 34-36 weeks, 0.81 (0.56-1.18); for 37-43 weeks, 1.21 (1.02-1.44).

| DISCUS SION
8][9][10] However, the use of aspirin for prevention of HDP has been debated because of contradictory findings among randomised trials, 21,22 and enthusiasm has been tempered by its relatively modest effect on the overall rate of reduction of HDP (RR 0.90, 95% CI 0.84-0.97). 5Several studies have suggested that the efficacy of LDA is contingent on early initiation and the use of higher doses (100-150 mg). 5,11,23More recently, a secondary analysis of the ASPRE trial has suggested that the true benefit of LDA may not be in preventing HDP but rather in delaying its onset to later gestational ages at which fetal/neonatal survival is more likely to occur. 12,13

| Main findings
This secondary analysis of the ASPIRIN trial similarly found that LDA did not decrease the overall rate of HDP at delivery but rather decreased the rate of HDP early in pregnancy. 17omen who received LDA and developed HDP appeared to deliver at a later GA as compared with women who received the placebo, suggesting that LDA may delay the onset of HDP at delivery.Specifically, we were able to document that among women with HDP, deliveries before 34 and 37 weeks were less common and directionally lower, although not statistically significant before 28 weeks.The lack of statistical significance before 28 weeks may reflect the small number of women who developed HDP at those gestational ages.Additionally, the probability of not having HDP at delivery was almost the same until 28 weeks GA.After 32 weeks GA, a dip in the trajectory was observed in the placebo group, F I G U R E 2 Gestational age at delivery for women with hypertensive disorders of pregnancy by treatment group.whereas at weeks GA, the probability of with HDP increased in the aspirin group.The calculated hazard ratios for early preterm birth were lower than for late preterm and term deliveries.
Similarly, the incidence of LBW, extreme LBW, fetal loss and stillbirths were significantly lower among women who received LDA than women who received placebo among the pregnancies complicated by HDP.Maternal outcomes such as antepartum haemorrhage, postpartum haemorrhage and vaginal bleeding did not differ between the two groups.

| Interpretation
Initiating LDA low-risk nulliparous singleton pregnancies early in gestation (6 +0 -13 +6 weeks) resulted in extending the gestational age at delivery among women who developed HDP.Similarly, the risk of perinatal mortality and SGA <10th percentile are reduced.These findings are consistent with other trials. 12,21,22Specifically, the finding that HDP women who received LDA tended to deliver at a later GA compared with women who received the placebo, is also consistent with a secondary analysis of the ASPRE trial that randomised women at high-risk for pre-eclampsia to LDA 150 mg or placebo, and found that the overall rate of HDP was not lowered but rather delayed to later gestational ages. 12,13It is interesting to note that the risk reduction of preterm delivery <34 weeks among women with HDP in this analysis (RR 0.37, 95% CI 0.17-0.81)was nearly identical to the risk reduction reported in the ASPRE trial (RR 0.38, 95% CI 0.20-0.70)despite a lower dose of aspirin being used. 12The risk of postpartum haemorrhage was not significantly increased but was higher among women with HDP receiving LDA.This increased risk has been reported in randomised trials targeting high-risk populations 12,24 and among studies evaluating universal aspirin prophylaxis in low-risk populations. 25,26

| Strengths and limitations
The strength of this investigation includes the large number of women who were recruited with negligible loss to followup, use of a placebo and masked ascertainment of the primary outcome.It should be remembered that these results F I G U R E 3 Estimated quantile of gestational age at delivery for women with hypertensive disorders of pregnancy by treatment arm for each quantile.
do not the results of the primary study, which was designed to examine the impact of LDA on preterm birth.
This was a non-prespecified analysis and in the primary paper study we did not explicitly ascertain whether HDP resulted in iatrogenic preterm delivery.Additionally, we had limited power to examine the outcome of HDP associated with preterm delivery before 28 weeks, a group that is particularly at risk for perinatal mortality in LMICs.

| CONCLUSION
Women randomised to LDA were less likely to have HDP at delivery before 34 and 37 weeks.Similarly, SGA <10th percentile and perinatal mortality occurs less frequently in those receiving LDA.This suggests that LDA improves pregnancy outcomes by delaying the onset of HDP, thereby extending the gestational age at delivery.

DATA AVA I L A BI L I T Y S TAT E M E N T
Deidentified participant data from the study will be available at the National Institute of Child Health and Human MKH, SSG, RJD, NG, JM, EMMcC, MK-T and RS designed the study.AK, MSS, MCM, SSG, AL, AT, MB, MM, EC, WAC, LF, NFK, SJ, SS, RLG, PD, AP, PLH, FE and SB collected data and oversaw the trial at each centre.MKH, SSG, RJD, NG, JM, EMMcC, RS and MK-T provided overall trial oversight.JM and EN statistically analysed the data.AK and MKH wrote the first draft of the paper.All authors reviewed and contributed to the article.F U N DI NG I N FOR M AT ION Eunice Kennedy Shriver National Institute of Child Health and Human Development, USA.

F I G U R E 4
Survival plot of hypertensive disorder of pregnancy by treatment group.
Maternal characteristics among women with hypertensive disorder of pregnancy.
T A B L E 1 Primary and secondary outcomes among women with hypertensive disorder of pregnancy.
T A B L E 2